A new approach to DNA sequencing is proposed. DNA primers synthesized to be complementary to specific sequences on targeted genes will be covalently tethered in known locations in an array on the surface of a glass slide and allowed to hybridize with and capture complementary target gene fragments. The primer/template duplexes will be serially interrogated by single species of fluorescently-labeled deoxyribonucleotide triphosphate (dNTP) in the presence of an exonuclease-deficient DNA polymerase. Those primer strands for which the template base adjacent to the 3' end of the primer is complementary to the base of the dNTP supplied will be extended by incorporation of one or several deoxynucleotide monophosphates; extension will cease, on all strands in an array spot, when the next template base is not complementary to the dNTP. Quantitative fluorescent imaging of the array will identify those primers that have extended and quantify the number of bases incorporated, thus reading a short length of sequence (one to several bases) at a known location on the target gene. The fluorescent label is then selectively destroyed and the cycle is continually repeated with all four types of dNTP. A manual reaction system developed in preliminary work will be automated for rapid cycling. Studies will be directed towards increasing read length and sequence accuracy by optimizing attachment chemistries and enzyme performance, reducing nucleotide impurities to negligible levels, calibrating any context-dependent in the fluorescence response, and correcting for signals arising from extension failure. Initially, the read length target per spot is at least 50 bases; primers will be spaced at short intervals (about 20-50 bases apart) along the target gene sequences so that a long sequence can rapidly be read out in short parallel bytes. Array densities > 10,000 spots are anticipated, with reaction cycle times approximately 2 min/dNTP giving data rates about 1,000 bases per minute on a single slide. Initial sequencing studies will address genes known to be associated with an elevated risk of cancer. As read length is increased, the technology will be applied to de novo sequencing by spotting cloned templates annealed to a universal primer corresponding to the vector sequence.